AI Paper Atlas

Research map / Label-efficient and robust learning / Semi-supervised learning

Semi-Supervised Learning for Optical Flow with Generative Adversarial Networks

Wei-Sheng Lai, Jia-Bin Huang, Ming-Hsuan Yang

NeurIPS 2017 · Label-efficient and robust learning › Semi-supervised learning · approach: GAN-based semi-supervised learning · 81 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add adversarial loss on flow warp error (lambda_adv 0.01) to supervised training (ablation)supervised SPyNet · KITTI 2012EPE: 7.69 -> 7.16Improved
add adversarial loss (lambda_adv 0.01) (ablation)supervised SPyNet · KITTI 2015EPE: 17.19 -> 16.02Improved
add adversarial loss (lambda_adv 0.01) (ablation)supervised SPyNet · Sintel-FinalEPE: 4.70 -> 4.68No clear effect
increase adversarial weight lambda_adv from 0.01 to 1 (ablation)proposed model · KITTI 2012EPE: 7.16 -> 13.89Hurt
use discriminator receptive field 47x47 instead of 23x23 (ablation)PatchGAN discriminator · KITTI 2012EPE: 7.38 -> 7.16Improved
use receptive field 47x47 instead of 95x95 (ablation)PatchGAN discriminator · KITTI 2012EPE: 7.54 -> 7.16Improved
add classical brightness and smoothness losses on unlabeled KITTI datasupervised SPyNet · KITTI 2012EPE: 7.69 -> 10.38Hurt
use adversarial semi-supervised learning instead of baseline semi-supervisedbaseline semi-supervised · KITTI 2015Fl: 39.33% -> 38.77%Improved
add unlabeled data with adversarial loss, 25% labeled datasupervised · KITTI 2012EPE: 10.60 -> 7.40Improved
add unlabeled data with adversarial loss, 25% labeled datasupervised · FlyingChairsEPE: 2.09 -> 2.33Hurt
add unlabeled data with adversarial loss, 25% labeled datasupervised · Sintel-CleanEPE: 4.91 -> 3.95Improved
vary baseline semi-supervised weights lambda_w and lambda_s (ablation)baseline semi-supervised · KITTI 2012EPE: 8.63 to 11.82Mixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy